Mercurial > repos > jjohnson > spectrast

comparison link_scan_datasets.py @ 5:274fdc50169b draft default tip

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/spectrast commit e5b5b15b0a995c8588ff62f92fd0a2329fb7a6a5-dirty
author jjohnson
date Wed, 25 Jul 2018 15:05:34 -0400
parents 7f02fc51bddf
children
comparison
equal deleted inserted replaced
4:c9bfe6adb7cd 5:274fdc50169b
1 #!/usr/bin/env python 1 #!/usr/bin/env python
2 2
3 from __future__ import print_function 3 from __future__ import print_function
4 4
5 import argparse 5 import argparse
6 import difflib
7 from difflib import SequenceMatcher
8 import os 6 import os
9 import os.path 7 import os.path
10 import sys
11 import xml.sax 8 import xml.sax
12 9 from difflib import SequenceMatcher
13 10
14 11
15 def __main__(): 12 def __main__():
16 parser = argparse.ArgumentParser( 13 parser = argparse.ArgumentParser(
17 description='link spectrum datasets to the name used' + 14 description='link spectrum datasets to the name used' +
18 ' in the identification dataset') 15 ' in the identification dataset')
19 parser.add_argument( 16 parser.add_argument(
20 'ident_files', nargs='+', 17 'ident_files', nargs='+',
21 help='Pepxml or mzIdentML') 18 help='Pepxml or mzIdentML')
22 parser.add_argument( 19 parser.add_argument(
23 '-n', '--scan_name', default=[], action='append', 20 '-n', '--scan_name', default=[], action='append',
24 help='Name for scan file') 21 help='Name for scan file')
25 parser.add_argument( 22 parser.add_argument(
26 '-f', '--scan_file', default=[], action='append', 23 '-f', '--scan_file', default=[], action='append',
27 help='Path for scan file') 24 help='Path for scan file')
28 args = parser.parse_args() 25 args = parser.parse_args()
29 26
30 class MzidHandler( xml.sax.ContentHandler): 27 class MzidHandler(xml.sax.ContentHandler):
28
31 def __init__(self): 29 def __init__(self):
32 xml.sax.ContentHandler.__init__(self) 30 xml.sax.ContentHandler.__init__(self)
33 self.spectraDataFiles = [] 31 self.spectraDataFiles = []
34 self.spectraDataNames = [] 32 self.spectraDataNames = []
35 self.searchDatabaseFiles = [] 33 self.searchDatabaseFiles = []
36 self.searchDatabaseNames = [] 34 self.searchDatabaseNames = []
35
37 def startElement(self, tag, attrs): 36 def startElement(self, tag, attrs):
38 if tag == 'SpectraData': 37 if tag == 'SpectraData':
39 id = attrs['id'] 38 id = attrs['id']
40 path = attrs['location'] 39 path = attrs['location']
41 filename = os.path.basename(path) 40 filename = os.path.basename(path)
49 filename = os.path.basename(path) 48 filename = os.path.basename(path)
50 name = attrs['name'] if 'name' in attrs else None 49 name = attrs['name'] if 'name' in attrs else None
51 self.searchDatabaseFiles.append(filename) 50 self.searchDatabaseFiles.append(filename)
52 self.searchDatabaseNames.append(name if name else id) 51 self.searchDatabaseNames.append(name if name else id)
53 print ("SearchDatabase: %s %s" % (name if name else id, path)) 52 print ("SearchDatabase: %s %s" % (name if name else id, path))
54 def endElement( self, name): 53
55 pass 54 def endElement(self, name):
56 def characters( self, data):
57 pass 55 pass
58 56
59 class PepXmlHandler( xml.sax.ContentHandler): 57 def characters(self, data):
58 pass
59
60 class PepXmlHandler(xml.sax.ContentHandler):
61
60 def __init__(self): 62 def __init__(self):
61 xml.sax.ContentHandler.__init__(self) 63 xml.sax.ContentHandler.__init__(self)
62 self.spectraDataFiles = [] 64 self.spectraDataFiles = []
63 self.spectraDataNames = [] 65 self.spectraDataNames = []
66
64 def startElement(self, tag, attrs): 67 def startElement(self, tag, attrs):
65 if tag == 'msms_run_summary': 68 if tag == 'msms_run_summary':
66 basename = attrs['base_name'] 69 basename = attrs['base_name']
67 name = os.path.basename(basename) 70 name = os.path.basename(basename)
68 ext = attrs['raw_data'] 71 ext = attrs['raw_data']
69 path = '%s%s' % (basename,ext) 72 path = '%s%s' % (basename, ext)
70 filename = os.path.basename(path) 73 filename = os.path.basename(path)
71 self.spectraDataFiles.append(filename) 74 self.spectraDataFiles.append(filename)
72 self.spectraDataNames.append(name) 75 self.spectraDataNames.append(name)
73 print ("SpectraData: %s %s" % (name, path)) 76 print ("SpectraData: %s %s" % (name, path))
74 def endElement( self, name): 77
78 def endElement(self, name):
75 pass 79 pass
76 def characters( self, data): 80
81 def characters(self, data):
77 pass 82 pass
78 83
79 parser = xml.sax.make_parser() 84 parser = xml.sax.make_parser()
80 parser.setFeature(xml.sax.handler.feature_namespaces, 0) 85 parser.setFeature(xml.sax.handler.feature_namespaces, 0)
81 handler = PepXmlHandler() 86 handler = PepXmlHandler()
82 parser.setContentHandler( handler ) 87 parser.setContentHandler(handler)
83 for ident in args.ident_files: 88 for ident in args.ident_files:
84 parser.parse(ident) 89 parser.parse(ident)
85 90
86 spectra_names = handler.spectraDataFiles 91 spectra_names = handler.spectraDataFiles
87 92
88 def best_match(name): 93 def best_match(name):
89 if name in spectra_names: 94 if name in spectra_names:
90 return name 95 return name
91 try: 96 try:
92 r = [SequenceMatcher(None, name, spectra_names[x]).ratio() for x in range(len(spectra_names))] 97 r = [SequenceMatcher(None, name, spectra_names[x]).ratio()
98 for x in range(len(spectra_names))]
93 return spectra_names[r.index(max(r))] 99 return spectra_names[r.index(max(r))]
94 except Exception, e: 100 except Exception, e:
95 print ("best_match: %s %s" % (name, e)) 101 print ("best_match: %s %s" % (name, e))
96 102
97 for i,name in enumerate(args.scan_name): 103 for i, name in enumerate(args.scan_name):
98 path = args.scan_file[i] if len(args.scan_file) > i else '' 104 path = args.scan_file[i] if len(args.scan_file) > i else ''
99 (root, ext) = os.path.splitext(name) 105 (root, ext) = os.path.splitext(name)
100 print ("SpectraFile: %s %s" % (name, path)) 106 print ("SpectraFile: %s %s" % (name, path))
101 iname = best_match(name) 107 iname = best_match(name)
102 print ("IdentName: %s %s" % (name, iname)) 108 print ("IdentName: %s %s" % (name, iname))
105 print ("%s -> %s" % (iname, path)) 111 print ("%s -> %s" % (iname, path))
106 112
107 113
108 if __name__ == "__main__": 114 if __name__ == "__main__":
109 __main__() 115 __main__()
110